Re: [Synoptic-L] A Mark-plus-Matthew thought experiment

Expand Messages
• ... Suppose we imagine a time when the gospels of Mark and Matthew had been written and circulated, but Luke had not yet been written. We might readily deduce
Message 1 of 63 , Sep 9, 2012
---------Ron wrote-----
Suppose we imagine a time when the gospels of Mark and Matthew had been
written and circulated, but Luke had not yet been written. We might readily
deduce that Matthew had been dependent on Mark. But what would we make of
the material unique to Matthew? Would we see behind it a single written
source? Would there be any trace of Q? Should 2ST supporters *expect* to be
able to reconstruct something like Q in those circumstances? If not, doesn't
this constitute evidence for the incoherence of Q?
-----------------------

Take a look at my 2011 NTS article: Is there Stylometric Evidence for Q?
[https://www.wiki.ed.ac.uk/display/DLM/DLM-stylometric-analysis has details]
and go to the section using Cluster Analysis.

This section is closest to the kind of experiment envisaged.

i.e. take a large quantity of sayings in Matthew _not_ derived from
Mark. Divide them into 250 word sections. Use only a few of the
words of the highest frequency in order to have results that produce
enough instances to be statistically viable. Use Cluster. See if Q
sayings and M sayings do, or do not, cluster into different groups.
Then pair up the samples to make 500 word samples, use the same criteria
and re-test (using larger numbers of these highest frequency words in
fewer samples). Then before making hasty remarks about sample sizes
(as is the wont of some) exercise some thought about how high the word
counts are which are derived from the samples.

As I say this experiment is close to the one suggested, but not
exact. I would be most interested to see if someone could design
an experiment that matched the exact requirements of Ron's very
interesting suggestion, but I would suggest giving the matter a
lengthy period of serious thought (i.e. months not days).

David M.

---------
David Mealand, University of Edinburgh

--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
• I have long been aware of general differences between the style of Luke and that of Acts. In 1986 Kenny drew attention to this issue in a new way, but there
Message 63 of 63 , Sep 15, 2012
I have long been aware of general differences between the style of
Luke and that of Acts. In 1986 Kenny drew attention to this issue
in a new way, but there still seemed to be no need to question common
authorship of the two texts. One obvious line, in addition to noting
similar themes, was that the earlier work was revising previous
sources written in a Greek more influenced by Aramaic than the Greek
of the author, even if the latter did sometimes like to imitate the
Septuagint. The later work is more concerned with Greece and Rome.
So the differences of style seemed explicable.

That situation changed with the publication of the monograph by
Patricia Walters, the reason being that this showed that the sections
of text most likely to be authorial in each of the works displayed
very significant stylistic differences. The criteria used were ones
known and used in the Graeco-Roman world. The results do not in
themselves require a decision for different authorship, but they
do require serious further explanation. They do call in question
a widespread assumption of common authorship, and demonstrate that
much more careful attention needs to be given to the issue if such
an assumption is to stand.

One obvious line is to check the new findings and this I did. I used
the same standard statistical method with different criteria. The
results were still highly significant. My criteria were very high
frequency, so I could partition the data further. The Luke sub-samples
cohered, the Acts sub-samples cohered, but the difference between the
two sets was still highly significant. I tried one further tactic
of removing some sections of the Luke seams and summaries to test a
further possible line of objection. The original results still stood.
I then used a different (multivariate) method and included other samples
from both texts to see where the SS samples landed in relation to those,
and to each other. I ran variations on that. One of those is on my
web site. After all these checks I was convinced that the original
statistical results were robust and needed to be taken very seriously.

I am not inferring directly from stats to different authorship. I am
concluding that my previous reasons (and those of others) for acquiescing
in the assumption of common authorship deserve serious reappraisal. A
really convincing explanation of the differences needs to be provided,
if such is in fact possible.

I am well aware that style can be affected by what an author admires or
detests, and by whether some disturbance of mood affects the composition
of a piece, or the author has an opium dream. While I have not checked
for the latter in the case of Luke-Acts, I have checked genre rigorously,
and most of the other items one might usually suspect. I also looked
at the lines of defence in as many of the scholarly reviews as I could
find. Obviously I am not going to post here all the items I would put
in an eventual properly published piece, and even there editors differ
considerably as to how extensively the data are presented. But I
would not be so definite about the need for reappraisal of the issue if
I had not carried out a lot of tests not immediately evident.

David M.

---------
David Mealand, University of Edinburgh

--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
Your message has been successfully submitted and would be delivered to recipients shortly.